Deconstructing geospatial agent-based model: Sensitivity analysis of forest insect infestation model

Abstract

Agent-based models (ABM) can be used to represent the spatio-temporal dynamics of real world geospatial phenomena, however because of their complexity, they can be difficult to implement and validate. This study uses the invariant-variant validation approach to further model testing of a developed ABM of forest insect infestation representing spatio-temporal dynamics of the emerald ash borer (EAB). The invariant-variant method deconstructs model results to facilitate an improved understanding of the model’s sensitivity to changes in input parameters and focuses on EAB agents’ access to information. Obtained results indicate that the developed EAB agent-based model represents and maintains both process accuracy and spatial similarity.